This project aims to increase our understanding of the etiology, development, heterogeneity and comorbidity of smoking by examining both individual and contextual influences. It focuses on adolescents, who are at greatest risk for smoking initiation (SI) and in whom prevention and control of smoking persistence (SP) would have the greatest potential benefit to public health. The last decade has seen tremendous growth in quantitative statistical methods for complex traits, and in large genetically informative epidemiological samples, which now allow us to test critical complex questions. We have access to three longitudinal genetically informative data sets of adolescent twins and their parents, six data sets with published adolescent smoking data and three extended twin (ET) kinship studies. Detailed data on smoking behavior are available in one adult and in three adolescent/young adult twin samples. Data are from the United States, Belgium, the Netherlands, Finland, Australia and Sweden. We will test the following hypotheses about SI: 1) SI in early adolescence is primarily influenced by shared environmental factors; genetic factors gradually contribute more to liability to SI in later adolescence and young adulthood, 2) the rates of SI vary by culture (US, Europe, Australia), but variance components do not, 3) the influence of parents and parental smoking decreases and that of siblings and peers increases from adolescence to young adulthood, 4) the influence of measured environmental factors gradually decreases and that of measured genes (identified in GWAS) increases over adolescence, 5) the effect of measured genes on SI varies as a function of environmental factors (GxE), 6) the role of genes and environment is the same for males and females. The second set of hypotheses concerns the relationship between SI and SP/ nicotine dependence (ND). We will test whether 7) there is significant but not complete genetic and environmental overlap between SI and measures of SP & ND in adolescence/ young adulthood, 8) there is significant measurement variance by age and sex for SP/ND, 9) the same genes and environments influence SI and SP throughout adolescence, 10) there is genetic overlap between SI, age of onset of SI and between age of onset and SP/ND, 11) the contribution of environmental risk factors is greater to SI than to SP/ND, and 12) the contribution of measured genes (identified in GWAS) is greater for SP/ND than for SI. Our access to rich developmental (prospective and retrospective) data sets and our expertise in statistical modeling of genetically informative samples, puts us in a unique and ideal position to address these specific aims. Doing so will substantially improve understanding of how nature and nurture interact to generate liability to smoking.